Neon lamp as a backemf Arrestor

[Mosaic]

Hi all:
I am considering using a neon lamp (or paralleled pair for current handling) rated at 65V ac trigger as an inductive kickback limiter in a 36VDC environment where the NFET switches appear to avalanche around 70+V. I see wiki mentions its use.
Does this sound reasonable?

 

[crutschow]

How high is the inductive current?
What is the inductor?

 

[ronsimpson]

Spark gap voltage limiters are used in TV sets for years.
I have seen neon (what looks like lamps) used on the video board.
I do not know what will happen if the lamp is hit hard and often. The lamps look pretty strong.
Please give it a try!

 

[Mosaic]

Current is maxed at about 700A. Inductor is perhaps 4uH max. dt for the current to drop to zero is 7 uS.
So energy in kickback =>1/2*LI*I = 1 Joule
V= L di/dt => 400V unchecked.

Edit...happening @ 500Hz!
The kickback is currently snubbed to around 60V peak at max energy via a 50SQ80 (2 bucks) pulse diode and a 1uF 63V PEN cap in series. Diode gets up to 55C or so. Without the snubber the voltage goes up to about 75V and avalanches thru the switching NFETs....
I was thinking a Neon lamp might be a more effective way ( and a lot cheaper @ 0.35c) and act as a display as well to confirm proper operation as I can't tell from casual observation that the snubber is working.

 

[crutschow]

700A?? I doubt that a neon bulb will handle that amount of current (at least not more than once or twice).

 

[Mosaic]

Um...no no,, the forward current charging the inductor is 700A. It collapses to zero in 7 uSecs thereby inducing a 400V kick back for about that long. It is the 400V kickback that the neon is supposed to clamp.
Further reading indicates that Neon lamps using a Penning mixture which ionizes at a lowish trigger voltage but takes some milliseconds to do so. They have been used to clamp transients in the past.

I guess I'm gonna have to test 'em...in some safe way in case exploding glass occurs!
I can test their switch on time easily enough I suppose, and if it's too slow I'll have to abandon the idea.

 

[rumpfy]

The short answer is NO.
The long answer is that the energy in the spike has to have its energy dissipated outside the switching device. A neon would not be able to conduct such energy unless it was VERY large. If the switching device is correctly selected, it will be able to withstand the energy.
The design of snubbing arrangements is generally quite complex because it is dealing with parasitic circuit elements. A 700 amp switch is quite a hefty thing and care is required.

 

[ccurtis]

I agree with crutschow that the bulb won't last very long. The reverse discharge current will be the same as the charging current, even if for a short period. If you look at gas discharge tubes designed for lightning protection, at 1000A they are only good for a few hundred (quick) shots.

The diode and cap in series (and I might include a small resistor too, to lower the Q) is actually not a bad idea, since the combination will limit the voltage by not allowing the energy (current) in the inductor to discharge so rapidly (V=L di/dt). A suitably rated TVS is also a possibility.

 

[JimB]

Also, do not forget that a neon tube needs a higher voltage to strike than it does to conduct.
So the peak voltage before the snubbing action starts will be higher that the conduction voltage of the neon tube.

JimB

 

[Scotophor]

... which is also why neons in parallel is a bad idea. The first one to conduct will lower the voltage, preventing those in parallel from conducting. At least, until it fails.

 

[crutschow]

Better to use a zener diode or transient suppressor diode to clamp the voltage spike.

 

[Mosaic]

The pulse diode /capacitor combo seems to work a LOT better than TVS in terms of longevity.

 

[crutschow]

The pulse diode /capacitor combo seems to work a LOT better than TVS in terms of longevity.

Whatever works best for you is good.